delayed, but not immediate recall.
Elzinga, B.M.; Bakker, B.; Bremner, J.D.
Citation
Elzinga, B. M., Bakker, B., & Bremner, J. D. (2005). Stress-induced cortisol elevations are
associated with impaired delayed, but not immediate recall. Psychiatry Research, 134,
211-223. Retrieved from https://hdl.handle.net/1887/14219
Version:
Not Applicable (or Unknown)
License:
Leiden University Non-exclusive license
Downloaded from:
https://hdl.handle.net/1887/14219
Stress-induced cortisol elevations are associated with impaired
delayed, but not immediate recall
Bernet M. Elzinga
a,T, Abraham Bakker
b, J. Douglas Bremner
c,d,ea
Section of Clinical and Health Psychology, University of Leiden, P.O. Box 9555, 2300 RB Leiden, The Netherlands
b
Robert-Fleury Stichting, National Center for Eating Disorders, Leidschendam, The Netherlands
c
Departments of Psychiatry and Behavioral Sciences and Radiology, Center for Positron Emission Tomography, Emory University School of Medicine, Atlanta, GA, USA
d
PET Center/Nuclear Medicine, Emory University Hospital, Atlanta, GA, USA
e
Atlanta VAMC, Decatur, GA, USA
Received 7 February 2004; received in revised form 26 July 2004; accepted 28 November 2004
Abstract
Glucocorticoids are known to modulate memory functions, with elevated cortisol levels being associated with impaired declarative memory. This specific effect has been shown in several studies using pharmacological doses of cortisol. The present study was designed to assess the effects of stress-induced cortisol elevations on (1) the type of memory processing (encoding, consolidation and retrieval), and (2) on the emotional valence of the material under study. Sixteen healthy females were presented neutral and emotional material (words and paragraphs) before and after a stress challenge. Declarative memory was tested immediately after presentation and 24 h later (delayed recall). Delayed, but not immediate recall of the information presented after the stress challenge was significantly reduced compared with delayed recall of information presented before the stress challenge. In line with this, strong negative correlations were found for delayed recall of words and spatial memory presented after the challenge with post-stress cortisol levels, whereas no significant correlations were found between cortisol levels and delayed recall at day 1. These results suggest that stress-induced cortisol specifically affects long-term consolidation of declarative memories. These findings may have implications for understanding the effects of traumatic stress on memory functioning in patients with stress-related psychiatric disorders.
D 2005 Elsevier Ireland Ltd. All rights reserved.
Keywords: Stress; Glucocorticoids; Memory; Hippocampus; Consolidation; Humans
1. Introduction
The relationship between stress and memory is currently a topic of considerable interest (Bremner, 1999; Lupien and Lepage, 2001; McGaugh, 2000; 0165-1781/$ - see front matterD 2005 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.psychres.2004.11.007
TCorresponding author. Tel.: +31 71 5273745; fax: +31 71 5274678.
Roozendaal, 2002; de Kloet et al., 1999). Evidence that stress or arousal can have important effects on memory functioning dates back to almost a century ago when the Yerkes–Dodson law was proposed describing an inverted U-shaped relationship
between arousal and memory performance (Yerkes
and Dodson, 1908). Moderate as compared with low levels of arousal facilitate learning and memory up to an optimal point, beyond which additional arousal leads to a successive decrease in memory functioning. Also from a clinical punt of view, there are clear indications that chronic stress results in persistent memory impairments, given the distur-bances in memory functions that are pervasive in stress-related psychiatric disorders, including depres-sion, posttraumatic stress disorder, and dissociative
disorders (Bremner et al., 1993; Elzinga and
Bremner, 2003; Burt et al., 1995; Dorahy, 2001). Hormones released during stress such as gluco-corticoids (GCs) and catecholamines have effects on memory that may provide a mechanism for the inverted U relationship between stress and memory. There are indications that catecholamines and low to moderate levels of GCs strengthen memory forma-tion in the short term (seeMcGaugh, 2000; Cahill et al., 2003), while high or chronic levels of GCs predominantly have an inhibitory effect on memory
function (see McEwen, 2000, Lupien and McEwen,
1997). In humans, GC receptors are widely
dis-tributed in the hippocampus and other brain regions,
including the prefrontal cortex (de Kloet et al.,
1999). GC receptors in the hippocampus are the site of negative feedback regulation of hypothalamic-pituitary-adrenocortical (HPA) function. The hippo-campus plays an important role in declarative memory, including the consolidation of short-term
into long-term declarative memory (Zola-Morgan
and Squire, 1990) and spatial memory (i.e., memory for dspatial representation of the environmentT; see
Maguire et al., 2000).
In humans, administration of high doses of GCs (dexamethasone, hydrocortisone, prednisolone) has generally been associated with impairments in declarative memory, although findings are mixed (for a review, see Wolf, 2003). In several studies, treatment with GCs selectively impaired performance in hippocampal-dependent forms of memory (declar-ative memory tasks, e.g., paragraph recall, cued
recall), leaving procedural memory (e.g., implicit
memory) unaffected (Kirschbaum et al., 1996;
New-comer et al., 1999), suggesting that cortisol interacts with hippocampal neurons to induce memory defi-cits. Besides acute actions, chronic effects of GCs on memory functioning have also been reported. For
example, Newcomer et al. (1994, 1999) found
impaired paragraph recall after a 4-day administra-tion of dexamethasone. Several studies did not find reduced declarative memory with GC administration
before the learning phase (Lupien et al., 1999; de
Quervain et al., 2000; Wolkowitz et al., 1990), although hydrocortisone did impair working memory (Lupien et al., 1999), and impaired delayed word
recall when administered 1 h before retrieval (de
Quervain et al., 2000), suggesting a role for prefrontal mediated memory impairments (see
Lupien and Lepage, 2001).
Pharmacological studies may not be representative of the physiological effects of endogenous cortisol release in humans exposed to natural stressors. Some exogenously administered drugs (e.g., dexametha-sone) are less able to penetrate the blood–brain barrier, and therefore they may not bind to
hippo-campal GC receptors (Lupien and McEwen, 1997).
Moreover, during exposure to stress, other stress hormones, including cathecolamines, are released at the same time. Recent evidence from animal studies suggests that these may interact with the effects of
GCs on memory functioning (Roozendaal, 2000). So
far, only a few studies have addressed the effects of stress-induced cortisol elevations on memory
per-formance in humans (Kirschbaum et al., 1996;
Lupien et al., 1997; Domes et al., 2002). In the
only study assessing a non-aged sample (Kirschbaum
et al., 1996), elevated cortisol levels induced by a psychosocial stress challenge were associated with poorer immediate recall of word pairs. Word recall was not assessed at baseline, however, so it remains unclear whether memory was actually impaired as a result of the stressful task. Moreover, no control tasks were administered so that the specificity of the cognitive impairments remained unclear. In a sample of healthy elderly participants, a stressful public speaking task induced a significant decrease in
learning and recall of word pairs (Lupien et al.,
middle-aged women who were either subjected to a public speaking task or a control condition. More-over, in a subgroup of high cortisol responders, memory performance even increased.
One factor that has often been ignored in the studies on stress-related memory functions is the impact of stress on the different types of memory. Memory processing involves at least three phases; acquisition, consolidation, and retrieval. At each phase, stress-related factors can come into play to affect memory formation. Most previous studies could not detect possible differential effects of GCs on distinct phases of encoding, consolidation, and retrieval, because both learning and recall were tested immediately after the stress challenge or, in the case of exogenous GCs, treatment affected both acquisition and retrieval.
Recent findings indicate that cortisol enhance-ment does not uniformly affect memory perform-ance for all information; rather it interacts with the emotional valence or degree of arousal at initial encoding of material in modulating memory for the material, presumably by interaction with
noradre-nergic activation (Buchanan and Lovallo, 2001;
Cahill et al., 2003; Roozendaal, 2000; Okuda et al., 2004). So far, two studies reported selectively enhanced delayed recall of emotionally arousing pictures, one after pre-learning cortisol administra-tion (Buchanan and Lovallo, 2001), the other after post-learning cortisol elevations induced by cold presser stress (Cahill et al., 2003). Rimmele et al. (2003) failed to replicate these findings, however. The present study was designed to assess the effects of stress-induced cortisol elevations on (1) the three types of memory processing (encoding, consolida-tion and retrieval), and (2) on the emoconsolida-tional valence of the material under study. Sixteen participants were presented neutral and emotional material (words and paragraphs) before and after a stress challenge. To differentiate between the effects of stress on the different phases of memory processing, half of the information was recalled shortly after presentation (i.e., to assess the effects of stress on encoding and retrieval), whereas the other half was recalled 24 h later (i.e., to assess the effects of stress on encoding and consolidation). Tests of declarative memory were used to assess the effects of stress on hippocampal functioning. Control tasks
of memory and attention were included to assess the specificity of the declarative memory impairments.
2. Methods 2.1. Participants
Sixteen female paid volunteers (Yale University students) with a mean age of 21.4 F 2.1 years participated in the study. They were recruited using local advertisements and were screened for general medical health and psychiatric disorders by a trained psychiatrist (A.B.) using the MINI International Neuropsychiatric Interview (Lecrubier et al., 1997). Participants were excluded for the presence of clinically significant medical illness or axis I psychi-atric disorder, including any substance abuse disorder based on the MINI. Participants had to refrain from strenuous physical exercise, large meals, cigarette smoking, and coffee for at least 1 h before the experiment because of the known effects of these variables on HPA functioning. After a brief introduc-tion to the study, all participants gave written informed consent for their participation in a protocol approved by the Human Investigation Committee of Yale University.
2.2. Stress challenge
2.3. Memory tasks 2.3.1. Word recall
Participants received two lists of 20 words, including 10 neutral words (e.g., dpicture,T dbuilding,T dpencilT), and 10 emotional words (e.g., dtortured,T dblood,T dscreamT). Emotional words were rated as more fearful, sad, nervous, angry, and less happy than neutral words (all P b 0.0001). Two parallel forms of the word lists consisting of different words compara-ble in difficulty were presented, balanced between baseline (before the challenge) and after the challenge (see Fig. 1). The first investigator (B.E.) read the words out loud. Participants were instructed to rate how dfearfulT each word was on a five-point scale. After the first presentation, the words were read a second time, and participants were instructed to rate how daversiveT each word was. This test was considered a test of incidental learning. To assess the effects of the stress challenge on encoding/ consolidation, recall of the words was tested 24 h later in two declarative, hippocampus-dependent memory tasks (cued recall and recognition task). An implicit memory task (word stem completion), which is assumed to be independent of hippocampal functioning, was included to test the specific effects of the stressor on declarative memory.
For the cued recall task, 20 target word stems were presented corresponding to 10 neutral and 10 emo-tional words presented at day 1. Half of these words had been presented before the challenge, and the other half had been presented after the challenge. Partic-ipants were asked to bfill in the word stems with words that were previously presented both before and after the stress challenge.Q
For the word stem completion task, the 20 remaining target word stems (not presented in the cued recall task) corresponding to 10 neutral and 10 emotional words presented at day 1, were presented intermixed with 20 filler word stems that did not correspond to any previously pre-sented words, to obscure the goal of memory testing. Half of the target words were presented before, and the other half were presented after the stress challenge. Participants were instructed to bfill in the first word that comes to your mind.Q The number of target completions served as a measure of implicit memory.
For the recognition task, the 20 (10 neutral and 10 emotional) words administered before and the 20 words administered after the challenge were pre-sented intermixed with 40 (20 neutral and 20 emotional) new words. Participants had to rate how confident they were that they had seen the word before at day 1 using the following rating scale: 1. I’m sure that the word is new (not studied before); 2. The word is probably new; 3. The word is probably old; 4. I’m sure that the word is old (studied before). Participants were given 2 points each time they assigned a d4T (sure that the word is old) to a target
word, 1 point for each d3,T 1 point for each d2,T
and 2 points for each d1.T The total score served as a measure of recognition memory.
2.3.2. Paragraph recall
The Wechsler Memory Scale-Revised Logical
Memory test (Wechsler, 1981) was used as a valid
and sensitive measure of verbal declarative memory that has proved to be sensitive to GC effects in
previous studies (Newcomer et al., 1994, 1999).
-Paragraph#1 IR -Distractibility -Working Memory
-Paragraph#2 E -Spatial#1 ER -Spatial#2 E
-Wordlist#1 -Paragraph#1 DR
-40 -30 -20 -10 0 10 20 30 40 50 60 (min)
s s s s s s s s s s s cognitive
challenge
-Paragraph#3 IR -Distractibility -Working Memory
-Paragraph#4 E -Spatial#3 ER -Spatial#2 E
-Wordlist#2 -Paragraph#3 DR
Fig. 1. Time line of memory tasks and saliva sampling.
Four paragraphs, each containing 25 pieces of information matched for difficulty, were constructed using an established method from the Wechsler Memory Scale-Revised Logical Memory. Two para-graphs were administered before the stress challenge serving as baseline measures, and two paragraphs were administered after the stress challenge. To distinguish the effects of stress on encoding, consolidation, and retrieval, two paragraphs were recalled immediately and after 30 min; the other two paragraphs were recalled only at day 2 (see
Fig. 1). Percent retention was computed as ddelayed recall / immediate recall 100%.T To avoid any non-random bias, presentation of the paragraphs was balanced over the four conditions, so that one paragraph was presented and tested before the challenge (baseline), one paragraph was presented before the challenge and tested at day 2, one paragraph was presented and tested after the challenge, and one paragraph was presented after the challenge and tested at day 2.
2.3.3. Spatial memory
The spatial memory test was a variation of a test
developed by Kirschbaum et al. (1996), which
proved to be sensitive to the effects of GCs. Participants were instructed to carefully read a short description of a walk in which they were dguidedT along a path with several dattractions,T e.g., specific trees, flowers, and animals that were situated either on the right or the left side of the path. Additionally, the stroller dsawT three bifurcations where she dturnedT left, right, or kept going straight ahead. Participants were given 3 min to memorize the description. Thereafter, they returned the description sheet to the investigator. For testing purposes, participants had to imagine that the stroller decided to return, walking back to the starting point of the walk. In a multiple choice test of 15 questions, participants had to choose the correct paths (at bifurcations) and describe whether an attraction was at the left or the right side of the path on the way back. Four parallel forms of descriptions of a walk were presented, so that for each participant (as in the paragraph recall) one story was presented and tested before the challenge, one story was presented before the challenge and tested at day 2, one story was presented and tested after the challenge, and
one was presented after the challenge and tested at day 2 (see Fig. 1).
2.3.4. Working memory
Working memory was measured using the digit recall subtest of the Wechsler Adult Intelligence
Scale-Revised (WAIS-R; Wechsler, 1987).
Partici-pants were presented two trials of six series of an increasing number of digits (from four to nine) that they had to repeat in the same order. In a second presentation, two trials of six series of digits (from three to eight digits) were presented that partic-ipants had to repeat in reversed order. This task was included to control for the specificity of the memory impairments, since working memory is not mediated by the hippocampus, but by the prefrontal cortex. Two parallel versions were presented in randomized order during baseline and after the challenge.
2.3.5. Continuous performance task
To control for attention, distractibility was mea-sured using a computer-generated continuous per-formance task that presented a rapid, continual sequence of numbers in one of three positions, right, center, or left (Gordon Diagnostic System, GDS;
Gordon and Mettelman, 1987). Participants were required to respond by pressing a key to each d9T that appeared immediately after a d1,T but only if they both appeared sequentially in the center position. The total number of correct responses and the reaction latencies were analyzed. The continuous performance task was administered during baseline and after the challenge.
2.4. Physiological assessment
Heart rate and blood pressure were assessed every 10 min using a Dynamap cuff (Critikon, parent company: GE Medical systems Informatics Technolo-gies, Milwaukee, WI). During the stress challenge, heart rate and blood pressure were assessed every 2 min. For the analysis of heart rate and blood pressure,
the means of the samples before ( 40, 30, 20, and
Cortisol samples were obtained during a 100-min
period at 11 assessment points, at 40, 30, 20,
10, 0, +10, +20, +30, +40, +50, and +60 min with reference to the start of the stressor. Cortisol levels were monitored using saliva samplings to avoid the stress-inducing effects of blood sampling. Determi-nation of cortisol in saliva provides a reliable measure of the free unbound fraction of cortisol (Tunn et al., 1992). In our laboratory, it was found that in subjects who had both plasma and salivary cortisol measures obtained simultaneously (N = 7), there was a high degree of correlation between these measures (r = 0.64, df = 6, P b 0.001). Saliva samples were collected using Salivette collection devices and
stored at 70 8C. Salivette tubes were centrifuged
(0 to 4 8C) to prepare saliva, which was analyzed for cortisol using a125I immunoradiometric assay kit available from Diagnostic Products Corporation (Los Angeles, CA). Samples and standards (200 Al) were determined in duplicate. For the analysis, repeated measures of the 11 cortisol samples were analyzed and cortisol peak with reference to baseline was computed by subtracting baseline (+0 min) and the cortisol peak at +40 min with reference to the beginning of the stressor (peak Cort). Day-to-day coefficients of variation for low (398 pg/ml) and high (4.12 ng/ml) concentration quality assessment samples were 10.1% and 8.4%, respectively. 2.5. Procedure
For the procedure, see Fig. 1. Because of
cortisol’s diurnal variations, testing was carried out between 1400 h and 1700 h. Participants were placed in a quiet room in a reclining chair with application of a Dynamap cuff for measurement of heart rate and blood pressure. After a resting period of 20 min, baseline memory testing was assessed
for 30 min in the following order (see Fig. 1):
paragraph#1 encoding and immediate recall;
para-graph#2 encoding only; encoding of neutral and
emotional words#1; distractibility task; spatial
mem-ory#1 encoding and recall; working memory; spatial
memory#2 encoding only; and paragraph#1 delayed
recall. After baseline memory testing, participants had a brief relaxation period of 5 min. Then the physician (A.B.) came in and carried out the stress challenge. After 20 min, the physician left the room,
and the participants had a 10-min relaxation period during which heart rate and blood pressure returned to baseline levels to minimize the acute (adrenergic) stress effects on memory. After the relaxation period, the second part of memory testing took place for 30 min in the same order as during
baseline (e.g., paragraph#3 encoding and immediate
recall; paragraph#4 encoding only; encoding of
neutral and emotional words#2; distractibility task;
spatial memory encoding and recall#3; working
memory; spatial memory#4 encoding only; and
paragraph#3 delayed recall). At the end of day 1,
participants were fully debriefed with respect to the purpose of the stress challenge (i.e., they were told that the negative feedback and the aversive attitude of the physician were fake). Participants were told that the purpose of the second appointment was to assess baseline memory functioning, and they were assured that no more stressful tests would be administered.
At day 2, participants came back for a 30-min surprise recall test. Testing took place in the
following order: recall paragraph#2; spatial
mem-ory#2; word stem completion (words#1+#2), cued
recall (words#1+#2), recognition (words#1+#2), spatial memory#4, and paragraph#4. Afterwards, participants were fully debriefed.
2.6. Analyses
3. Results
3.1. Physiological measurements
The cognitive challenge resulted in an increased heart rate (78.13 F 3.47 bpm) relative to baseline (68.82 F 2.55 bpm) and recovery (68.65 F 2.88 bpm) (main effect for time: F2,14= 13.01, P b 0.001; post hoc pairwise comparisons with Bonferroni correction of dchallenge versus baseline,T P b 0.0001; dchallenge versus recovery,T P b 0.0001; dbaseline versus recov-ery,T P = NS). The cognitive challenge also resulted in an increased systolic blood pressure (115.51 F 2.73 mmHg), relative to baseline (105.28 F 2.33 mmHg) and recovery (106.32 F 1.86 mmHg) (main effect for time: F2,14= 10.36, P b 0.01; pairwise comparisons of dchallenge versus baseline,T P b 0.001; dchallenge versus recovery,T P b 0.001; dbaseline versus recovery,T P = NS). This was the same for diastolic blood
pressure ( F2,14= 24.51, P b 0.0001), with a mean
diastolic blood pressure of 73.09 F 1.27 mmHg during challenge; 63.83 F 1.46 mmHg during baseline; and 65.22 F 1.57 mmHg during recovery (pairwise com-parisons of dchallenge versus baseline,T P b 0.0001; dchallenge versus recovery,T P b 0.0001; dbaseline versus recovery,T P = NS). Analysis of the cortisol levels resulted in a cubic interaction (F1, 11= 5.70,
P b 0.05). The increase between the lowest point before the challenge (t0, 10.6 F 1.2 Ag/dl) and the peak cortisol response after the challenge (t40, 15.0 F 2.8 Ag/dl) was 41.5% ( F1,15= 4.32, P b 0.05, 1-tailed); seeFig. 2.
3.2. Memory
For memory performance on day 1, seeTable 1; for day 2, seeTable 2.
3.2.1. Wechsler memory task
When tested 24 h later, participants tended to recall fewer items from the paragraph presented after exposure to the stress challenge than before. No differences were found on day 1 between
immediate, delayed, and percentage recall
(delayed / immediate recall 100) before and after the stress challenge.
3.2.2. Spatial memory
When tested 24 h later, participants made more errors in the questions on the descriptions of walks that were presented after the stress challenge than before. No differences were found between perform-ance on the spatial task on day 1 before and after exposure to the stress challenge.
8 9 10 11 12 13 14 15 16 17 -40 -30 -20 -10 0 10 20 30 40 50 60 Time (minutes)
Free cortisol in saliva (nmol/l)
3.2.3. Words
3.2.3.1. Cued recall. When tested 24 h later, words that were presented before exposure to the stress challenge were somewhat better recalled than words presented after the stress challenge, but in the cued recall task, this difference did not reach signifi-cance. When tested separately, recall of neutral words was impaired by the stress challenge, whereas recall of emotional words was not. Overall, emotional words were better recalled than neutral words ( F1,15= 10.10, P b 0.01).
3.2.3.2. Recognition. Words that were presented
before the stress challenge were better recognized than
words presented after the stress challenge. Overall, emotional words were better recognized than neutral words ( F1,15= 19.00, P b 0.001). Moreover, recogni-tion of neutral words presented after the challenge was relatively more impaired than recognition of emotional words, but this interaction between valence and time was not significant ( F1,15= 2.47, P = NS).
3.2.3.3. Word stem completion. Words presented
before exposure to the stress challenge tended to be more often completed than words presented after the stress task, but this effect was not significant. Neutral words were completed as often as emotional words ( F1,15= 0, P = 1). No differences were found between neutral and emotional word stem
comple-Table 2
Memory performance at day 2 on spatial memory, Wechsler memory, cued recall, word stem completion and recognition of information presented at day 1 before and after the challenge (mean F S.E.M.)
Memory task Baseline After challenge F-value P-value
Mean F S.E.M. Mean F S.E.M. Two-tailed
Spatial memory 12.38 F 0.56#2 9.94 F 0.76#4 6.67 0.02
Paragraph recall 10.28 F 0.70#2 7.53 F 1.56#4 3.20 0.09
Cued recall (total) 1.09 F 0.17#1 0.84 F 0.14#2 1.71 0.21
Neutral words 0.81F0.21 0.31F 0.15 3.75 0.07
Emotional words 1.38 F 0.15 1.38 F 0.29 0.00 1.0
Word stem completion (total) 0.33 F 0.08#1 0.17 F 0.06#2 3.95 0.07
Neutral words 0.31F 0.15 0.19F 0.10 0.48 0.50
Emotional words 0.34 F 0.12 0.16 F 0.09 1.22 0.29
Recognition (total) 14.06 F 1.07#1 9.06 F 1.44#2 35.19 0.000
Neutral words 11.06 F 1.73 4.25 F 2.18 16.23 0.001
Emotional words 17.06 F 1.12 13.89 F 1.41 8.34 0.011
#1corresponds to words#1;#2corresponds to paragraph#2and spatial memory#2, words#2;#4corresponds to spatial memory#4and paragraph#4
(seeFig. 1). Table 1
Memory performance at day 1 on spatial memory, Wechsler memory, working memory, and distractibility task during baseline and after the challenge (mean F S.E.M.)
Memory task Baseline After challenge F-value P-value
Mean F S.E.M. Mean F S.E.M. Two-tailed
Spatial memory 14.00 F 0.30#1 13.75 F 0.36#3 0.60 0.45 Paragraph recall Immediate recall 18.97 F 0.74#1 18.19 F 0.95#3 0.52 0.48 Delayed recall 17.28 F 0.76#1 16.56 F 1.03#3 0.40 0.54 % retention (delayed/immediate) 91.14 F 1.96#1 90.62 F 2.63#3 0.03 0.86 Working memory 17.63 F 1.09 19.00 F 0.74 6.15 0.03 Distractibility task # correct answers 27.06 F1.12 28.50 F0.93 0.76 0.40 Reaction time 45.56 F 1.89 46.25 F 1.85 0.19 0.67
tion before and after the stress challenge ( F1,15= 0.40, P = NS).
3.2.3.4. Distractibility task. No differences were
found in distractibility before exposure than after exposure to the stress challenge in reaction latency or number of errors.
3.2.3.5. Working memory. Participants repeated more digits correctly after exposure to the stress challenge than before the stress challenge.
3.3. Relationship between cortisol and memory performance
To further evaluate the relationship between cortisol and memory functioning, Pearson correlations were calculated. To calculate correlations between cortisol levels and memory performance at day 1 and post-stress cortisol levels during encoding and delayed memory performance at day 2, task performance was correlated with the absolute cortisol levels, based on the salivette that was closest in time to the task at hand (30 min post stress level for paragraph recall, 40 min for word recall, and 50 min for spatial memory). Correlational analyses on day 1 immediately before and after the stress challenge between cortisol levels
and memory performance on Wechsler immediate and delayed recall, spatial memory and working memory did not yield any significant correlation (all P N 0.10). At day 2, strong negative correlations were found between cortisol levels during encoding and cued word recall at day 2 (see Table 3 andFig. 3), as well as a
Table 3
Pearson correlations between delayed recall at day 2 of information learned before the stress challenge and peak cortisol (40 min post stress level) and between information learned after the stress challenge and post-stress cortisol levels during encoding Memory tests at day 2 Rbefore Rafter
Paragraph recall 0.08#2 0.02#4
Spatial memory 0.06#2 0.46#4
Cued recall total 0.21#1 0.76#2***
Neutral 0.45#1 0.23#2
Emotional 0.31#1 0.62#2**
Word stem completion 0.44#1 0.22#2
Neutral 0.27#1 0.13#2
Emotional 0.36#1 0.13#2
Recognition 0.18#1 0.37#2
Neutral 0.20#1 0.23#2
Emotional 0.03#1 0.09#2
Rbefore=Pearson correlation between peak cortisol level and delayed
recall of information learned before the stress challenge. Rafter=Pearson correlation between cortisol levels at encoding and
delayed recall of information learned after the stress challenge. ** P b 0.01. *** P b 0.001. cortisol (µg/100 ml) .3 .2 .1 0.0 4 3 2 1 0
Number of correctly recalled words
r = -.76
negative correlation between spatial memory and cortisol levels, which went up to r = 0.71, P b 0.01 one cortisol measurement later, after 60 min (seeTable 3). To evaluate the potential effects of stress-induced cortisol elevations on memory consolidation of infor-mation presented before the stress challenge, correla-tions were calculated between the peak cortisol response to the stress challenge (at +40 min) and delayed recall of information presented before the challenge. Except for the word stem completion task, all correlations were small and in a positive direction (seeTable 3).
In addition to the association of memory perfor-mance with absolute cortisol levels, the association with net increases of cortisol was also investigated. These analyses yielded similar results, i.e., negative correlations with post-stress net increase and spatial memory and explicit word recall at day 2, and no significant correlations on any other task at day 2 or day 1.
4. Discussion
The results of this study suggest that stress exposure may specifically affect long-term memory consolidation, as we found a reduction in the delayed recall of information presented after a stressful event compared with the delayed recall of information presented before stress exposure. In line with these findings, delayed recall of (emotional) words and spatial information learned after the stress challenge was negatively correlated with cortisol levels during encoding, whereas correlations with delayed recall of information presented before the stress challenge were all non-significant. Interestingly, exposure to the stress challenge did not affect memory perform-ance immediately after the challenge. Consistent with these findings, no significant correlations were found between cortisol levels and memory performance at day 1, either before or immediately after the stress challenge. In line with previous studies, declarative memory (i.e., recognition, spatial memory, and paragraph recall) was especially affected by the stress challenge, whereas performance on hippo-campus-independent tasks was unaffected (i.e., dis-tractibility), or even improved after the challenge (i.e., working memory).
The observation that exposure to the stress challenge selectively affected delayed recall of infor-mation without directly affecting performance imme-diately after acquisition is consistent with the genomic actions of GCs, as few GC actions in the hippocampus are executed until about an hour after the onset of the
stressor (see McGaugh, 2000; Lupien and McEwen,
1997). The crucial question is what specific processes may have mediated the decrease in delayed recall of information presented after the stress challenge relative to information administered before the stres-sor? Given the fact that cortisol levels remained elevated after cessation of the stressor for 30 min, stress-induced cortisol increases may potentially have affected both acquisition and consolidation of infor-mation presented after the challenge, while for information learned before the stress challenge, only consolidation can have been influenced. Following this line of reasoning, the pattern of result can be interpreted as both (i) impaired acquisition of infor-mation learned after the stressor, and (ii) enhanced consolidation of information learned before the stress challenge.
Pertaining to the first interpretation, the significant negative correlations between cortisol and delayed recall of information learned after the stressor would argue for a role of GC-related impaired acquisition. Moreover, GC-induced impaired acquisition would be in line with previous findings of decreased word recall (immediately) after stress-induced cortisol increases (see Kirschbaum et al., 1996; Lupien et al., 1997),
although findings are mixed (see Wolf, 2003, for a
consolidation of information learned before the stress challenge. Although pre-clinical evidence for GC-induced enhanced memory consolidation is compel-ling, direct empirical support is weak, given the small correlations that have been found between cortisol levels and delayed recall of information learned before the challenge. Correlational data on cortisol and memory consolidation should be interpreted with caution, however, given the fact that memory con-solidation is a process of hours or even days, and the salivary cortisol sample represents only a relatively arbitrary fraction of the total amount of GCs a person is exposed to during this period. In conclusion, the significant negative correlations between cortisol and delayed recall suggest a role for GC-induced impaired acquisition. Future studies are clearly needed to disentangle the effects of stress exposure on acquis-ition and consolidation.
Irrespective of the phase of learning, recall of emotionally arousing material was enhanced relative to neutral information. Extensive animal research suggests that enhanced memory for emotional material is related to an interaction between stress hormones (e.g., cortisol and epinephrine) and the degree of arousal at initial encoding of the material to be learned (Roozendaal, 2000). The association between cortisol and memory for neutral versus emotionally arousing material has not yet been systematically explored in humans, however. So far, two studies reported selective enhanced delayed recall of emotionally arousing pictures, one after pre-learning cortisol
administration (Buchanan and Lovallo, 2001), the
other after post-learning cortisol elevations induced by cold presser stress (Cahill et al., 2003).Rimmele et al. (2003) failed to replicate these findings, however. Despite the enhanced recall of emotional material, we could not detect significant positive correlations between cortisol levels and the recall of emotional words presented either before or after the stress challenge, as may be expected if cortisol activity is causally related to enhanced memory consolidation of emotional information. Instead, we found a strong negative correlation between cortisol levels and recall of words learned after the challenge, which was strongest for emotional words. One explanation for this discrepancy might be that words are not arousing enough to induce strong emotion. This issue needs to be explored in future studies.
Our failure to find immediate effects of GC-induced cortisol levels on memory performance is in
line with two recent studies (Wolf et al., 2001;
Domes et al., 2002), but is in contrast with two earlier studies showing memory impairments
imme-diately after stress exposure (Kirschbaum et al.,
1996; Lupien et al., 1997) or after cortisol
admin-istration (Newcomer et al., 1994, 1999; Kirschbaum
et al., 1996). Several factors may account for these discrepant findings. Participants in our study con-sisted of young female participants, whereas the samples in the studies showing immediate memory impairments after stress consisted either of elderly (Lupien et al., 1997) or male participants ( Kirsch-baum et al., 1996). Cortisol increases are known to have a stronger (negative) impact on memory functioning in elderly than in a young population, presumably because baseline cortisol levels are higher among the elderly. Second, consistent gender differ-ences in cortisol responses to psychological stress have been reported, with females generally showing smaller cortisol responses than males and less pronounced associations between cortisol and
mem-ory performance (see Kirschbaum et al., 1999). For
example, only men showed clear cortisol elevations after a stress challenge in the study ofKirschbaum et al. (1996), and negative correlations between cortisol and immediate memory recall were found only in men (Wolf et al., 2001). In a similar vein, no immediate memory impairments appeared in a study in which only women participated (Domes et al., 2002). Third, compared with pharmacological studies, peak cortisol levels in response to the cognitive challenge were moderate in size (41%) and far below those obtained in pharmacological studies (e.g., Lupien et al., 1999; Newcomer et al., 1999). Given the fact that the cognitive challenge was clearly perceived as stressful in terms of subjective distress, heart rate, and blood pressure, the moderate cortisol increase may in part be related to the fact that only females participated. Pharmacological studies have demonstrated a dose– response curve of cortisol and memory performance, showing that low doses did not impair memory
performance (Newcomer et al., 1999; Lupien et al.,
1999). It could be speculated that more pronounced
Several limitations of the study need to be kept in mind when evaluating the present findings. Most importantly, the study did not include a control condition without a stressor. It is therefore difficult if not impossible to disentangle the effects of the stressor from other potential factors like interfer-ence, practice effects (which may have played a role in the enhanced working memory), or changes in motivation. In our opinion, the pattern of selective interference with delayed declarative recall, and not with immediate recall, working memory, and atten-tion, argues against a major influence of these non-specific factors on memory performance. Future studies examining memory performance in a stress and a control condition, including a male and a female sample, are needed to elucidate the effects of stress on memory performance. Second, the correla-tional data in the present study should be interpreted with caution, given the increased risk of false positives with multiple comparisons. Finally, there are limitations in not controlling for oral contra-ceptives and menstrual cycle, factors that may modulate memory performance in women.
To conclude, the results of this study suggest that increases in stress-related cortisol may specifically affect delayed recall of stress-related experiences. To our knowledge, this is the first study that has directly compared immediate and delayed effects of cortisol on memory functioning. Given the genomic effects of GCs that may take hours or even days, studying the delayed effects of GCs may be a viable way to understand the relation between stress and memory functioning. To study the relevance of the present findings for traumatized individuals, it would be interesting to assess the effects of endogenous GC levels and other stress-related hormones and transmitters on (distinct phases of) memory functioning in patients with stress-related disorders, such as depression or PTSD. Also, studies are needed that assess GC levels during and immediately after traumatic events, and during the course of the development of PTSD in
association with memory changes (see also
Sapol-sky, 2000). These studies can make an important contribution to our understanding of the complex effects of stress on memory, and can be of potential help in the treatment and prevention of memory disturbances in stress-related disorders.
Acknowledgments
This study was supported by a grant from the Foundation for Behavioral and Educational Sciences of the Netherlands Organization of Scientific Research (NWO) awarded to Prof. dr. van Dyck (#575-29-003) and by NIMH R01 MH56120, VA Career Development Award, and the National Center for PTSD Grant awarded to Dr. Bremner.
The authors thank Helen Sayward, M.S., Kathy West, B.S., and Monica Lee Carter for assistance in conducting these studies.
References
Bremner, J.D., 1999. Does stress damage the brain? Biological Psychiatry 45, 797 – 805.
Bremner, J.D., Scott, T.M., Delaney, R.C., Southwick, S.M., Mason, J.W., Johnson, D.R., Innis, R.B., McCarthy, G., Charney, D.S., 1993. Deficits in short-term memory in posttraumatic stress disorder. American Journal of Psychiatry 150, 1015 – 1019. Bremner, J.D., Vythilingam, M., Vermetten, E., Adil, J., Khan, S.,
Nazeer, A., Afzal, N., Elzinga, B., Schmahl, C., McGlashan, T., Anderson, G., Heninger, G., Southwick, S.M., Charney, D.S., 2003. Cortisol response to a cognitive stress challenge in posttraumatic stress disorder (PTSD) related to childhood abuse. Psychoneuroendocrinology 28, 733 – 750.
Buchanan, T.W., Lovallo, W.R., 2001. Enhanced memory for emotional material following stress-level cortisol treatment in humans. Psychoneuroendocrinology 26, 307 – 317.
Burt, D.B., Zembar, M.J., Niederehe, G., 1995. Depression and memory impairments: a meta-analysis of the associa-tion, its pattern, and specificity. Psychological Bulletin 117, 285 – 305.
Cahill, L., Gorski, L., Le, K., 2003. Enhanced human memory consolidation with post-learning stress: interaction with the degree of arousal at encoding. Learning and Memory 10, 270 – 274.
de Kloet, E.R., Oitzl, M.S., JoJls, M., 1999. Stress and cognition: are corticosteroids good or bad guys? Trends in Neurosciences 22, 422 – 426.
de Quervain, D.J.F., Roozendaal, B., Nitsch, R.M., McGaugh, J.L., Hock, C., 2000. Acute cortisone administration impairs retrieval of long-term declarative memory in humans. Nature Neuro-science 3, 313 – 314.
Domes, G., Heinrichs, M., Reichwald, U., Hautzinger, M., 2002. Hypothalamic–pituitary–adrenal axis reactivity to psychological stress and memory in middle-aged women: high responders exhibit enhanced declarative memory performance. Psychoneur-oendocrinology 27, 843.
Elzinga, B.M., Bremner, J.D., 2003. Are the neural substrates of memory the final common pathway in PTSD? Journal of Affective Disorders 70, 1 – 17.
Gordon, M., Mettelman, B.B., 1987. Technical Guide to the Gordon Diagnostic System. Gordon Systems, Inc., Syracuse, NY. Kirschbaum, C., Wolf, O.T., May, M., Wippich, W., Hellhammer,
D.H., 1996. Stress- and treatment-induced elevations of cortisol levels associated with impaired declarative memory in healthy adults. Life Sciences 58, 1475 – 1483.
Kirschbaum, C., Kudielka, B.M., Gaab, J., Schommer, N.C., Hell-hammer, D.H., 1999. Impact of gender, menstrual cycle phase, and oral contraceptives on the activity of the hypothalamus– pituitary–adrenal axis. Psychosomatic Medicine 61, 154 – 162. Lecrubier, Y., Sheehan, D.V., Weiller, E., 1997. MINI International
Neuropsychiatric Interview: a short diagnostic structured inter-view. Reliability and validity according to the CIDI. European Psychiatry 5, 224 – 231.
Lupien, S.J., Lepage, M., 2001. Stress, memory, and the hippo-campus: can’t live with it, can’t live without it. Behavioral Brain Research 127, 137 – 158.
Lupien, S.J., McEwen, B.S., 1997. The acute effects of corticoids on cognition: integration of animal and human model studies. Brain Research Reviews 24, 1 – 27.
Lupien, S.J., Gaudreau, S., Tchiteya, B.M., Maheu, F., Sharma, S., Nair, N.P.V., Hauger, R.L., McEwen, B.S., Meaney, M.J., 1997. Stress-induced declarative memory impairment in healthy elderly participants: relationship to cortisol reactivity. Journal of Clinical Endocrinology and Metabolism 82, 2070 – 2075. Lupien, S.J., Gillin, C.J., Hauger, R.L., 1999. Working memory is
more sensitive than declarative memory to the acute effects of corticosteroids: a dose–response study in humans. Behavioral Neuroscience 113, 420 – 430.
Maguire, E.A., Gadian, D.G., Johnsrude, I.S., Good, C.D., Ashburner, J., Frackowiak, R.S., Frith, 2000. Navigation-related structural changes in the hippocampi of taxi drivers. Proceed-ings of the National Academy of Sciences of the United States of America 97, 4398 – 4403.
McEwen, B.S., 2000. The neurobiology of stress: from serendipity to clinical relevance. Brain Research 886, 172 – 189.
McGaugh, J.L., 2000. Memory – a century of consolidation. Science 287, 248 – 251.
Newcomer, J.W., Craft, S., Hershey, T., Askins, K., Bardgett, M.E., 1994. Glucocorticoid-induced impairment in declarative mem-ory performance in adult humans. Journal of Neuroscience 14, 2047 – 2053.
Newcomer, J.W., Selke, G., Melson, A.K., Hershey, T., Craft, S., Richards, K., Alderson, A.C., 1999. Decreased memory performance in healthy humans induced by stress level cortisol treatment. Archives of General Psychiatry 56, 527 – 533.
Okuda, S., Roozendaal, B., McGaugh, J.L., 2004. Glucocorticoid effects on object recognition memory require training-associated emotional arousal. Proceedings of the National Academy of Sciences of the United States of America 101, 853 – 858. Rimmele, U., Domes, G., Mathiak, K., Hautzinger, M., 2003.
Cortisol has different effects on human memory for emotional and neutral stimuli. NeuroReport 14, 2485 – 2488.
Roozendaal, B., 2000. Glucocorticoids and the regulation of memory consolidation. Psychoneuroendocrinology 25, 213 – 238. Roozendaal, B., 2002. Stress and memory: opposing effects of
glucocorticoids on memory consolidation and memory retrieval. Neurobiology of Learning and Memory 78, 578 – 595. Sapolsky, R.M., 2000. Glucocorticoids and hippocampal atrophy in
neuropsychiatric disorders. Archives of General Psychiatry 57, 925 – 935.
Seeman, T.E., Berkman, L.F., Gulanski, B.I., Greenspan, S.L., Charpentier, P., Rowe, J.W., 1995a. Self esteem and neuro-endocrine response to challenge: MacArthur studies of success-ful aging. Journal of Psychosomatic Research 39, 69 – 84. Seeman, T.E., Singer, B., Charpentier, P., 1995b. Gender differences
in patterns of HPA axis response to challenge: MacArthur studies of successful aging. Psychoneuroendocrinology 20, 711 – 725.
Tunn, S., Molmann, H., Barth, J., Derendorf, H., Krieg, M., 1992. Simultaneous measurement of cortisol in serum and saliva after different forms of cortisol administration. Clinical Chemistry 38, 1491 – 1494.
Wechsler, D., 1981. The Wechsler Adult Intelligence Scale-Revised. Harcourt Brace Jovanovich, Inc., New York.
Wechsler, D., 1987. Wechsler Memory Scale-Revised. Harcourt Brace Jovanovich, Inc., New York.
Wolf, O.T., 2003. HPA axis and memory. Best Practice & Research in Clinical Endocrinology & Metabolism 17, 287 – 299. Wolf, O.T., Schommer, N.C., Hellhammer, D.H., McEwen, B.S.,
Kirschbaum, C., 2001. The relationship between stress induced cortisol levels and memory differs between men and women. Psychoneuroendocrinology 26, 711 – 720.
Wolkowitz, O.M., Reus, V.I., Weingartner, H., Thompson, K., Breier, A., Doran, A., Rubinow, D., Pickar, D., 1990. Cognitive effects of corticosteroids. American Journal of Psychiatry 147, 1297 – 1303.
Yerkes, R.M., Dodson, J.D., 1908. The relation of strength of stimulus to rapidity of habit formation. Journal of Compre-hensive Neurology and Psychology 18, 459 – 482.
